WriteHumanly
Free · No sign-up · Runs in your browser

Signs of AI Writing Checker

Paste a draft and see the patterns that make writing read as AI-generated, highlighted where they appear, with a plain explanation and a fix for each one.

Highlighted
Signs of AI writing are highlighted here as you type.

What the checker looks for

Staging instead of stating

Contrasts that reject something nobody said ("not just a tool, it's a movement"), run-ups like "let's dive in", closing lines that repeat the paragraph, and rebuttals to objections nobody raised.

Inflated significance

Ordinary facts described as pivotal moments, lasting legacies or testaments to something larger, plus unnamed experts standing in for evidence.

Overused AI words

Words models reach for far more often than people do: delve, tapestry, crucial, pivotal, vibrant, showcase, underscore and others. One is fine; a cluster stands out.

Formatting by rule

Em dashes joining every other clause, and bullet lists where every item opens with a bold label and a colon.

Chat leftovers

"Great question!", "I hope this helps", "Let me know if you'd like more", and notes about a model's knowledge cutoff left in finished text.

Patterns, not proof

The patterns come from Wikipedia's Signs of AI writing guide, written by the editors who review AI-generated articles on Wikipedia, and the open-source humanizer checklist built on it. They describe the default choices a language model makes. A person can make any one of them on purpose, so a single highlight means very little. Many highlights in one passage are worth an edit.

This tool does not estimate whether a text was written by AI, and it should never be the basis for accusing anyone. For a detector-style reading, try the free AI detector. To fix formatting debris like em dashes and hidden characters in one click, use the AI text cleaner. To rewrite a whole draft so it sounds natural while keeping your facts, names and numbers, use the WriteHumanly humanizer.

Frequently asked questions

Is this an AI detector?

No. It does not give a probability that text was written by AI. It highlights specific writing patterns that readers and editors associate with AI text, explains each one and suggests a fix. For a detector-style score, use the free WriteHumanly AI detector.

Where do these patterns come from?

From Wikipedia's "Signs of AI writing" guide, maintained by editors in WikiProject AI Cleanup, and the open-source humanizer checklist built on it. We turned the patterns that can be checked reliably by rule into this tool.

Does a highlight mean the text was written by AI?

No. Careful human writers use some of these patterns on purpose, and one sighting proves nothing. Treat the highlights as editing suggestions, and never accuse someone of using AI based on a checklist.

What is the difference between strong and weak signs?

Strong signs, such as chatbot leftovers or a "not just X, it's Y" contrast, are worth fixing on their own. Weak signs, such as a single em dash or the word "crucial", only matter when many appear together.

Is my text uploaded anywhere?

No. The checker runs entirely in your browser. Nothing you paste is sent to a server or stored.